The goal of this proposed research is to generate a genetic architecture of brain development. We aim to achieve this goal in two solid steps. We first utilize our existing brain MR imaging measurements (traditional MRI volumetry, brain connectivity, resting state activity) of 235 individuals, 125 with autism and 110 typically developing subjects, to build a data structure of brain regions of interest that have at least 50% estimated heritability for typical brains, sorted first by size and then by components of structura-functional variability. In our next step, we interrogate whole-exome and whole-genome sequences and network analysis of molecular pathways by state-of-the-art molecular genetics methods in reference to this data structure. We expect that this detailed analysis of variability and extremes in brain imaging phenotypes in autism will result in the discovery of new genetic associations that reveal new biological pathways and advance our understanding of autism neuropathology and its severely impairing clinical manifestations. Our results will thus change imaging genetics in autism research.